Learning Forum

[kabit]

Hello everyone,

in order to calculate the mass exchanged across a surface for a steady state simulation, I have created a seperated mesh region. In the post processing, I splitted the region into a surface with a positive normal velocity component and a surface with a negative velocity component. The results of the massFlow function for both surfaces are approximately zero, which should not be the case. Calculating the mass flow by areaint(u*density) gives reasonable results. However, the results for the two surfaces should be equal, but differ significantly.

Does anybody know, why this happens and if there is a workaround for this problem?

 

[Rob]

If you're reporting on an interior surface turn it into a wall (you'll see wall & wall:shadow) and then back to an interior.  It's because the interior surface normals aren't set in the mesh. 

[kabit]

Thank you very much for your reply.

Do you know if there is also a way to do this in CFX?

[Rob]

Probably much the same approach in CFD Post using the velocity components normal to the surface(s) or a flux. 

[DrAmine]

You can try using the mesh location for that. For that reason you  decompose in pre-processing and then define monitor in CFX-Pre and access the primitive 2D boundary in CFD-Post.

[kabit]

Thank you both for your help.

The calculation of the mass flow on the interior surfaces seems to work now. However, another problem has now occured. If I calculate the mass flow across an iso-clip surface, I get different results than on the interior surface, even if both surfaces are equal.

Does anybody know, how to fix this?

[Rob]

How different? The isosurface will pull data from surrounding cells and interpolate, the interior calculates the value directly with no interpolation. Also remember the issue with normals I mentioned. 

[kabit]

Hi,

sorry for my late respond. I just used the interior surface directly for my calculations to avoid the interpolation errors, which seems to work fine.

Thank you very much for your help.